Reprints Available Directly from the Publisher Photocopying Permitted by License Only Thymic Microenvironment and Lymphoid Responses to Sublethal Irradiation

Sublethal irradiation of the murine thymus has been a useful tool for depleting the thymus of dividing immature thymocyte subsets, to sequence thymocyte differentiation events occurring from radiation-resistant precursors. This massive reduction in thymocytes also represents a model in which the bidirectional interplay between the thymic stromal cells and lymphocytes can be investigated. The purpose of this study was thus twofold: to precisely map the initiation of thymopoiesis as a prelude to assessing the effects of injected mAb to novel thymic antigens; and to use a panel of mAbs to determine the alterations in the thymic stroma during the T-cell depletion and reconstitution phases. The striking finding from this study was that following T-cell depletion, there was a marked upregu-lation of specific stromal antigens, which retracted with the reappearance of T cells. Thus, following sublethal irradiation, there are modifications in the thymic microenvironment that may be necessary to support renewed thymopoiesis and the complete restoration of the thymus involved the synchronous development of both the stromal and lymphocytic components.


INTRODUCTION
The thymic microenvironment consists of specialized cells that are both epithelial and nonepithelial in nature. Together these cellular elements are organized into well-defined cortical and medullary regions, throughout which developing thymocytes reside . As T cells mature within the thymus, they migrate from the cortex to the medulla and undergo phenotypic changes, including the acquisition of TcR, cytokine receptors, and the modulation of differentiation antigens such as CD4 and CD8 (Boyd and Hugo, 1991). The contribution of the thymic microenvironment toward this complex process, and that which thymocytes themselves impart on the thymic microenvironment, is gradually becoming defined. Thymic stromal cells such as TNC, Mq, and DC form complexes with developing thymocytes (Andrews and Boyd, 1985;Kyewski, 1986;van Ewijk, 1988;Shortman and Vremec, 1991;Gao et al., 1993) and stromal cell lines instruct some, but not all, specific stages of Corresponding author. thymocyte development (Palacios et al., 1989;Nishimura et al., 1990;Tatsumi et al., 1990;Nagamine et al., 1991;Hugo et al., 1992;Watanabe et al., 1992;Anderson et al., 1993). Accordingly, damage to the microenvironment itself via drug treatment or ionizing irradiation impairs thymocyte differentiation (Adkins et al., 1988;Kanariou et al., 1989).
That the constitution and organization of the thymic microenvironment are in turn dependent on T-cell development has been well illustrated in a number of experimental systems. In SCID mice and in mice deficient in RAG-1 or p56 lck, the medulla fails to develop due to the absence of mature TcR thymocytes (Shores et al., 1991;Molina et al., 1992;Suhr et al., 1992;Ritter and Boyd, 1993;van Ewijk et al., 1994). Similarly, in TcR knockout mice, where thymocyte development is blocked beyond CD4/CD8 cells, the medulla is greatly reduced in size (Philpott et al., 1992;Palmer et al., 1993;Mombaerts et al., manuscript in preparation). This effect has also been demonstrated in mice that have been injected with anti-CD3 mAb (Kyewski, 1991). Furthermore, a thymocyte-stromal cell signaling pathway in which medullary epithelial cells are activated via the phosphorylation of a 90-kD 102 E.S. RANDLE-BARRETT and R.L. BOYD medullary epithelial-cell glycoprotein upon contact with CD4+CD8 thymocytes has been identified (Couture et al., 1992). Thus, it is evident that a bidirectional relationship exists between thymocytes and stromal cells (reviewed by Ritter and Boyd, 1993).
In the present study, we have mapped thymocyte reconstitution post-irradiation (PIrr) to ultimately devize a model for examining the in vivo effects specific mAbs have on thymopoiesis. Using a panel of mAb-recognizing mouse thymic stromal (MTS) antigens derived in this laboratory (Godfrey et al., 1990;Tucek et al., 1992), we also observed the effects of irradiation and the associated loss of T cells on the thymic microenvironment. In agreement with others (Huiskamp et al., 1983;Adkins et al., 1988), the initial loss of lymphocytes following irradiation resulted in the collapse of the stromal architecture, in particular, that of the cortex. Most importantly, however, it was clearly illustrated that the thymic was "dynamic" immediately following irradiation, with marked upregulation in the expression of specific cortical and medullary stromal antigens. These selective changes in the thymic stroma may have been as a direct result of the irradiation itself, revealed as a consequence of the rapid depletion of thymocytes or a prerequisite for the initiation of thymopoiesis following irradiation.

Effect of Irradiation on Thymocyte Subpopulations
Total Thymocyte Cell Number. Total cell yields dramatically decreased immediately following sublethal irradiation (Fig. 1) from their normal adult value (day 0) of 6.0+1.4x107 to the nadir of 4.2+ 1.2 x 105, by day 5. Between days 6 and 7, however, there was extensive thymocyte proliferation as indicated by the tenfold increase in total cell number to 4.4+0.31x 106, with an additional tenfold increase in number by day 12 to 3.3 +0.2x 107. CD3, CD4, and CD8 Expression. To assess thymocyte subpopulations and their subsequent reconstitution, thymocytes on days 1-14, 21, and 28 PIrr were stained simultaneously for CD3, CD4, and CD8 antigen expression. By day 1, the thymus is virtually devoid of CD4/CD8 double-positive (DP) thymo- Call (day 0) mice and sublethally irradiated mice, 1-28 days postirradiation. Results represent the mean + standard deviation of four independent experiments, except for days 8, 9, and 13, which are represented as the mean of two experiments. cytes (Fig. 2), the small remnant population being mainly of the CD4/CD8/CD3 hi subset (Fig. 3). The resistance of mature populations to the irradiation was further reflected in the mature (CD3/) CD4 and CD8 single-positive populations. Although these cells markedly increased in proportion within the first 4 days (Figs. 2 and 3), they progressively decreased in number (from 4.3+1.0x106 to 7.5+0.7x104 and 1.2+_0.4x106 to 1.8+_3.0x104, respectively) until day 10, presumably due to emigration (K. Kelly, personal communication). Similarly, CD3+CD4-CD8 thymocytes were unaffected proportionally (Fig. 3) but decreased in number from 3.0_+0.8x105 to 1.0+1.0x104 by day 5.
Thymic reconstitution began around day 4 with sequential, progressive increases in CD3-CD4- CD4+CD8 cells. Thus, between days 5-7 following irradiation, all immature thymocyte populations markedly expanded. By day 7, the thymocyte profile with respect to CD3, CD4, and CD8 expression approximated that of the adult thymus ( Fig. 2), although it is not clear whether the mature singlepositive cells were remnant radiation-resistant cells that had not yet migrated or were derived from de novo production of differentiating precursors. From day 7, all thymocyte populations increased in cell number, and by day 21, normal adult values were virtually attained (data not shown).
MTS 32, 33, 35 and 37. MTS 32, 33, 35, and 37 identify antigens with the unusual property of being expressed on both thymocytes and thymic stromal cells, thereby providing an additional phenotypic profile of thymocytes distinct from that defined by CD3, CD4, and CD8 expression (Godfrey et al., 1990;Tucek et al., 1992). On thymocytes, MTS 32 stains all T cells, except a subset of CD3hiCD4 CD8-cells, which exhibit a Th0/Th2-type cytokine profile (Vacari et al., 1994); MTS 33 and 37, which detect ThB and HSA, respectively, stain immature thymocytes and mature thymocytes, but are negative on peripheral T cells; MTS 35, detects the antigen TSA-1/Sca-2, which is restricted to immature thymocytes and has an inverse expression to CD3 . Figure  increased to approximately the adult values by day 12. In accordance with the dramatic loss of immature thymocytes, PIrr, ThB, TSA-1/Sca-2, and HSA expression were markedly reduced from approximately 90% to 2-5% by day 4, despite the fact that HSA and ThB are expressed on mature CD3 cells, albeit at low levels, which are virtually the only cells remaining after irradiation.

Microenvironment
The effects of sublethal irradiation on the thymic microenvironment were examined using an extensive panel of mAbs reactive with epithelial and nonepithelial stromal cells and the vasculature, the normal adult thymus reactivities have been described elsewhere (Godfrey et al., 1990). The data are summarized in Tables 1 to 4. Epithelial Cells. Following irradiation, the loss of rapidly dividing thymocytes and decrease in thymus size were associated with a collapse of the thymic stromal network, particularly that of the cortex. This was well illustrated by anti-CD4 and anticytokeratin labeling of the thymic epithelium (Fig.  5). Similarly, MTS 44 is pan-cortical epithelium, but also weakly stains infrequent, isolated medullary epithelial cells in the normal adult thymus. By 3 days PIrr, MTS 44 epithelial cells formed a compact 106    PIrr, the expressions of MTS 32 had vanished, leaving weak isolated epithelial cell staining in the cortex. Pockets of granular staining (possibly dead/ dying thymocytes) were also observed within epithelial cell-free regions. From days 3-6, MTS 32 remained weakly localized to infrequent epithelial cells and thymocytes within the cortex; reactivity had normalized by day 12. MTS 33 (ThB), normally reactive with cortical thymocytes and medullary epithelial clusters, displayed no thymocyte reactivity, but the latter were more obvious from days 1-6 PIrr, presumably as a result of the absence of T cells. Additionally, isolated clusters of weak ThB epithelial cells were revealed in the subcapsular cortex (Fig. 10). In accordance with flow cytometry, MTS 35 (TSA-1/Sca-2) and MTS 37 (HSA) were all negative on thymocytes from days 1-5, their reactivity on both cortical and medullary thymic stromal-cell subsets being more exposed. By day 6, their staining patterns had normalized.

DISCUSSION
In view of the recent realization that a major feature of thymic organogenisis is the bidirectional, symbiotic development relationship between thymic lymphocytes and stromal cells, the purpose of this study was to examine the responsiveness of these stromal cells to a dramatic loss in T cells. That is, has the microenvironment encompassed within the stromal elements the flexibility to alter in response to the need to reinitiate or enhance T-cell differentiation?
If so, can such a study reveal which specific stromal molecules are involved in this important process?
In addition to the anticipated dramatic loss of thymocytes, the most striking feature of the present study was the enhanced staining of selective thymic stromal-cell antigens following sublethal irradiation. There are three plausible explanations for this. Antigenic determinants, normally camouflaged by thymocytes, may have simply been exposed due to the loss of lymphocytes; the stromal cells may be responding directly to the effects of the irradiation; the sudden loss of thymocytes may have caused an upregulation in the stromal cell expression of these antigens. In accordance with clearly established data of previous studies (Takada et al., 1969;Huiskamp et al., 1983;, following irradiation there was indeed a dramatic decrease in thymic mass and collapse of the cortical stromal network through the loss of immature thymocytes. The developmental status of the remnant and regenerating T cells was verified not only by the changes in CD3, CD4, and CD8 profiles and the collapsed cortex, but also by the loss of TSA-1/Sca-2 (MTS 35), ThB (MTS 33), and HSA (MTS 37) positive cells. MTS 35 detects the antigen TSA-1/Sca-2, which is expressed only on immature thymocytes and is normally absent on peripheral T cells . Similar to HSA, ThB is present on most thymocytes, including the mature (CD3/) CD4 and CD8 siffgle-positive populations, but is absent on peripheral T cells . By days 3-4 PIrr, there was virtually no staining for ThB and HSA. This was surprising as phenotypically mature thymocytes were still present. It is possible that these antigens are radiosensitive or are cleaved by macrophagereleased proteases, but this would infer that the stromal cells that are still stained by the mAbs express different, more stable conformations of the antigens. There is no evidence to suggest immigration of mature peripheral (HSA-and ThBnegative) T cells into the irradiated thymus. The apparent buildup of CD3/CD4-CD8 and CD3/CD4 CD8 thymocytes presumably indicates a lag between phenotypic and functional (at least in terms of migration ability) maturity or it reveals a population of cells that may never leave the thymus reflecting the very low migration rate (,,1%/day; Scollay et al., 1980). Given the dramatic loss of cortical thymocytes, this could in principle expose the stromal antigens and hence the appearance of increased expression. This in part may be an explanation, but it does not account for the selective nature of the upregulated antigens and in any case would only be relevant in the cortex, but several antigens were enhanced in the medulla (MTS 9,14,16,17,and 29). From a purely technical viewpoint, it is also unlikely that the lymphocytes would normally sterically mask stromal antigens because they are nonoverlapping independent cell types and examined in the same plane of the section. Direct irradiation induced enhancement of the stromal antigens is also unlikely because we have observed similar, albeit less pronounced, increases following thymocyte depletion with hydrocortisone and 5-fluorouracil (unpub-lished observations). We thus favor the hypothesis that although some of the increases may be part of the "housekeeping" repair mechanisms following architectural damage to the thymus, they are essentially a direct response of the stromal-cell subsets to specifically reestablish a microenvironment geared to reinitiating or elevating basal levels of T-cell differentiation. Indeed, such specific upregulation of these molecules may be a generic feature of the thymus undergoing elevated thymopoiesis because we have found similar selective upregulation of stromal antigens during early ontogenesis (E14) and in the postcastration reversal of thymic atrophy (Price et al., manuscript in preparation).
Within the first 2 days following irradiation, TN thymocytes did not alter significantly, indicating their relative resistance to the irradiation; as to whether there were subtle shifts in the CD44/ CD25/c-kit defined TN subsets was not determined.
It is from within these TN cells that reconstitution begins around day 4 is rapidly expanded by day 7 and is virtually complete by day 21. High levels of [3H] TdR incorporation (Kadish and Basch, 1975;Sharp and Thomas, 1975) and increased lymphocyte reactivity of the mAb MTS 35,32,37,and ThB also indicated this to be a time of active cell division and increased immature thymocyte content.
Based on the findings reported herein, a synopsis of the intrathymic events following sublethal irradiation would be as follows. The initial and probably sole direct effect of the irradiation is the immediate loss of immature thymocytes, excluding the nondividing TN precursors. This causes an associated collapse of the cortex, the epithelium of which is still present, but as a compact zone of MTS 44 cells surrounding the medulla. From 24 to 72 h, there was a dramatic increase in macrophages and/ or myeloid cells revealed by enhanced MAC-l, MTS 17, MTS 37, and MTS 28 (data not shown) staining, presumably in response to the need to remove the dead and dying thymocytes. This increase in phagocytic cells is in agreement with earlier reports (Duijvestijn et al., 1982;. The loss of thymocytes was also associated with a marked increase in expression of antigens associated with the vasculature. Because it is unlikely that significant angiogenesis occurred during this time, the apparent increase in vascular endothelium (MTS 12 was more likely due to it being exposed. There was, however, an upregulation in the expression of the secreted endothelial antigen detected by MTS 15 and the extracellular matrix associated with the vasculature (MTS 16). Although the latter may be part of a tissue rebuilding process, we have also observed such endothelial activation in the regenerating male mouse thymus following castration (Price et al., manuscript in preparation). Hence as TN thymic precursors are localized around the thymic endothelium, it is possible that the vasculature initiates their proliferation and/or differentiation, evident from day 4 PIrr. The dynamic nature of the thymic stromal cells was further demonstrated by the dramatic increase in MTS 9 expression, which virtually stained the entire thymus from 24-72 h and gradually retracted with the increase in newly generated T cells, to the predominantly medullary epithelium pattern by day 6.
Hence, despite the widely held view that thymic stromal cells are a sessile population, they are capable of rapid alteration at both the cellular and molecular levels in response to the sudden depletion of thymocytes or need for renewed thymopoiesis. The post-irradiation model provides a valuable means of revealing this and we are currently using it to test the functional significance of the antigens by injecting the appropriate purified mAbs.

Animals
CBA/CaH male mice 4-6 weeks of age were obtained from the Monash University Central Animal House. The mice were exposed to 7.5 Gy of whole body y irradiation at a dose rate of 0.3 Gy/min, at the Walter and Eliza Hall Institute using Eldorado 6 teletherapy unit (Atomic Energy of Canada, Commercial Products) charged with 5000 Ci of 6Co.
Mice were subsequently killed at 12 h, days 1-14, 21, and 28 after the irradiation. Results are based on a minimum of four independent experiments and a pool of one to six mice per experiment, depending on the time point.

Cell-Surface Staining
All labeling for CD3, CD4, and CD8 was performed using anti-CD3-FITC (145-2Cll), anti-CD4 phycoerythrin (PE) (GK1.5), and anti-CD8-Biotin (B) (53-6.7) (CD4 and CD8 both from Becton Dickinson, CA). For three-color labeling, streptavidinphycoerythrin Texas red conjugate (Tandem TM, Southern Biotech, AL) was used to detect the THYMIC RECONSTITUTION AFTER IRRADIATION 115 biotinylated mAb. To control for labeling efficiency, freshly prepared thymocyte suspensions from normal adult mice were routinely used. For MTS mAb thymocyte reactivity, cells were indirectly labeled with the desired mAb followed by FITC-conjugated sheep anti-rat Ig (Silenus Laboratories). Stained cells were monitored using a FACScan (Becton Dickinson). Dead cells and nonlymphoid cells were excluded from data acquisition on the basis of 0 and 90 scatter profiles. Analyses were performed using Lysys II research software (Becton Dickinson).

Immunohistology
Thymii were snap frozen on liquid nitrogen and 4 m sections were cut using a cryostat. MTS mAb (Godfrey et al., 1990) labeling and the coexpression of epithelial cell determinants were assessed by double-labeling sections with the desired mAb and a polyvalent rabbit anti-cytokeratin Ab (Dako, Carpinteria, CA), respectively. Bound mAb was revealed by FITC-conjugated sheep anti-rat Ig (Silenus Laboratories) and TRITC-conjugated goat antirabbit Ig (Silenus Laboratories). By using these species combinations, no intermediate blocking steps were necessary. For three-color immunofluorescence, sections were labeled with anti-CD4-FITC, anti-CD8-Biotin (both from Becton Dickinson, and the polyvalent rabbit anti-cytokeratin Ab, washed and stained with a mixture of streptavidinphycoerythrin Texas red conjugate (Tandem TM, Southern Biotech) and AMCA anti-rabbit Ig (Jackson Immunodiagnostics, WG). Stained sections were washed and mounted under a coverslip using veronal buffered glycerol (pH 8.6) and examined using a Zeiss Axioskop fluorescence microscope. All photography was performed using Kodak 1000 ASA print or 1600 slide films.